U . S.

Department of Commerce Research Paper RP2105

National Bureau of Standards Volume 44, June 1950

Port of the Journal of Research of the Notional Bureau of Standards

Measurement of Sixty-Degree Specular Gloss

By Harry K. Hammond, III, and Isadore Nimeroff
Specular gloss is the attribute next to color that is most often used in the evalua(,ion
of the appearance of obj ect s. For 10 years t he American Society for T estin g N[a(,erials
has used a method of test for the 60 0 specular gloss of pain t fini shes based on resea rch done
at this Bureau. The ASTM method prescribes the illuminator and receiver apertu res that
are to be used. However, in order to determine the uncertain ties involved in the calibra-
tion of gloss standards, the rate of change of gloss reading with change of aperture was
investigated throughout the gloss scale. Thi s was done by modifying an existing gonio-
photometer so that it became in effect a versatile glossmeter with provis ion for accuratcly
controllin g all of the known geometrical vari ab les involved. Data obtained on the var ia-
tion of gloss readin g with aperture show that standards with widely different distr ibutions
of re flected ligh t can be used to check t he adjustment of glossmeters and the conformance
of thei r apertures to ASTM specifications in the in creasingly impor tan t medium- to high-
glos range. The problem of obtaining the d ir-ru e correction, which in vol ves sou ree polari za-
tion and depa rt ure of the specimen from a pe rfect diffu so r, has been investigated.

r. Introduction
The appearance of an object depends upon
several factors; the illuminant, the reflection char-
acteristics of the material , the surface texture, the
illuminating and viewing geometry, and the
observer . The color and glossiness of a specimen
are determined by the spectral composition and
geometrical distribution of the incident light and
upon the transformations that take place upon
reflection from the specimen. These two appear-
ance attributes, color and gloss, arc somewhat
related. Color measurement may require cor-
rection for surface glossiness. Specular gloss
measurement may require correction for diffuse
reflectance.
The complete description of the light-reflecting
properties of a specimen would require an infinite
number of measurements of reflectance throughout
a hemisphere for an infinite number of directions
of illumination. Even an approach to such a
set of data is uneconomical to obtain with present
goniophotometric equipment. Usually one must
be content to obtain one goniophotometric reflec-
tion curve for a single direction of illumination
(see fig. 1). Furthermore, the question of how to
interpret a gon iophotometric curve still remains,
for in contrast to spectrophotometric curves there
ANGLE OF T A N G L E OF
INCIDENCE SPECULAR
a REFLECTION
a'
PERPENDICULAR
PAINT SPECI MEN
FIGURE 1. T ypical goniophotometric curve for a paint
spe6men , with defining terms.
has not yet been developed a satisfactory method
for reducing the former to a simple meaningful
numeric or set of numerics as has been done for
the latter. B ecause of these circumstances, one
must generally be content to m easure the specular
reflectance for one direction of illumination.
Previous investigations have shown that the
choice of direction as well as aperture for illumina-
tion and view is governed by the type of material
involved. Hunter and Judd [1]1 found that 60°
specular gloss correlated most satisfactorily with
1 Figures ill brackets indicate the literature references at the end of this
paper.

Measurement -of 60° Specular Gloss 585

 the surface appearance of paint finishes, and a
method based on their investigations at this
Bureau was tentatively adopted by ASTM in
1939. In order to simplify the discussions in this
paper the problems treated are confined to 60°
specular gloss, but the considerations involved
apply to other specular angles as well.
Specular gloss could be measured without in-
strumental complication if there existed (1) an
infinitesimal, high intensity, unpolarized light
source; (2) perfect collimators; and (3) an in-
finitesimal but sensitive receiver. Then gloss-
meters could be readily duplicated. Practical
sources, collimators, and receivers cause difficulty
because imperfections in their construction com-
bine with differences in reflected-light-flux distri-
butions to give inconsistent gloss readings for
various types of materials. If it were practical to
make glossmeters with identical geometrical fea-
tures, there would be no difficulty in comparing
measurements made on different instruments. A
more practical approach to the problem, however,
lies in further standardizing gloss measurement by
specifying tolerances for glossmeter geometry.
The original investigation, already referred to,
on which the ASTM method is based, employed
rectangular source and receiver apertures. This
was done for two reasons. First, the type of
lamp employed in the apparatus had a single,
coiled filament producing a long narrow source,
the cross section of which is rectangular; second,
if the coiled lamp-filament were oriented with its
axis perpendicular to the plane of measurement
(see definition below) more light would be available
because a larger aperture could be permitted in a
plane perpendicular to the plane of measurement
without appreciably affecting the gloss reading.
Similarly, a proportionately larger receiver aper-
ture in the plane perpendicular to the plane of
measurement permits the receiver to gather light
that would otherwise be lost because of change in
direction on reflection from brush marks on
brushed paint panels, when these are oriented, as
they should be, with the marks parallel to the
plane of measurement.
In order to determine the permissible geometri-
cal tolerances, a "versatile" glossmeter has been
developed. This instrument has provision for
making the following adjustments:
586
l
1. The receiver aperture may be varied either
parallel to or perpendicular to the plane of
measurement.
2. The souree aperture may be varied either
parallel to or perpendicular to the plane of meas-
urement.
3. The position of the image of the source may
be varied with respect to the plane of the receiver
entrance window.
4. The specular angle may be varied.
In addition to determining the permissible
geometrical tolerances, there remains the problem
of obtaining the diffuse correction. Measured
values of diffuse correction do not agree exactly
with those obtained by simple theory; however,
the small discrepancy is accounted for by (1)
The departure of the reflected light distribution
of the specimen from that of the perfect diffusor;
(2) The degree of polarization of the source; and
(3) The reflectance of the surface.
II. Terms, Symbols, and Definitions
The purpose of glossimetry is to evaluate ob-
jectively some variable that may be used as an
index of appearance ascribable to gloss. The
most conveniently used variable involves some
function of the geometrical distributions of the
reflected light. For this reason it is advisable to
define common reflectance terms as well as the
more specialized concepts associated primarily
with gloss measurement.
In addition to the definitions given below, there
are two general concepts that should be discussed.
The adjective "luminous" should be used when-
ever it is important that the defined term include
weighting of the spectral energy distribution ac-
cording to the luminosity function . A distinction
must also be drawn between what may be called
surface reflectance and body reflectance. Re-
flection of the latter type may be described as re-
emission when the light penetrates the surface of
the specimen and re-emerges at the incident face
with a geometric distribution similar to that of a
self-luminous source. The definitions of specular
reflectance and diffuse reflectance given below
involve the concepts of surface reflectance and
body reflectance (re-emission), respectively.
Journal of Research
 The most important terms, definitions, and
symbols used in this paper are given b elow.
Illuminance. Luminous flux incident per unit
area of surface .
Luminous intensity (of r eflected luminous flux in
any direc tion). The solid-angular flux density.
Luminance. Luminous flux emitted per unit solid
angle and p er unit proj ect ed area.
Luminous (total ) reflecti on. A general term for
the process by which a part of th e incident
luminous flux leaves a specimen from th e inci-
dent side.
Specular angle, (J'. The angle b etween the per-
p endicular to a plane surface and th e r efl ect ed
ray. This angle is numerically equal to the
angle of inciden ce, (J , and lies in the same plane
on the opposite side of th e perp endicular; it is
often designated -(J (refer t o fig. 1).
Lumi nous sp ecular reflecti on. The process by
which incident luminous flux is r efl ected from a
surface in an image-forming state.
Lumi nous diffuse reflection . The process by which
incident luminous flux is reemitted (in a non-
image-forming state) .
Luminous specular reflectance. R atio of luminous
fiux leaving a surface in an image-forming state
to the inciden t luminous flux .
Luminous diffuse reflectance . Ratio of luminous
flux reemitted at the illuminated urface of a
nonself-Iuminous specimen to the inciden t
luminous flux.
Luminous (total ) reflectance. R atio Of luminous
flux reflect ed by a specimen to the luminous
flux incident on it. Diffuse r efl ectance plus
specular r efl ectance equals total r eflec tance.
It should b e noted that the specular and diffuse
r efl ectance components often cannot be sepa-
rated for m easurement.
Luminous directi onal refl ectance . Ratio of lumi-
nous flux from the specimen when illuminated
and viewed from specified directions to that
from the perfectly reflecting , perfectly diffusing
specimen, similarly illuminated and viewed.
Luminous fracti onal reflectan ce of a specimen.
Ratio of luminous flux reflect ed wit-hin a sp ecified
solid angle to the luminous flux incident on the
specimen .
Goni ophotometer. An instrument used to obtain
the geometric distribution of luminous flux .
Measurement of 60 ° Specular Gloss
Goni ophotometric reflection curve . A plot of the
geometric distribution of r eflect ed luminous
flux, for a specified distribution of inciden t flux,
as a function of angle of view (see fig. 1) .
Gloss . The function of luminous directional r e-
flectance of a specimen responsible for its shiny
or lustrous appearance.
Sp ecular gloss, Gs . The ratio of t hc specularly
r eflected luminous flux , F x , from a specimen
surface for a specified angle of incidence, (J, to the
specularly r eflected luminous flux, F ., from a
standard surface for the same angle of incidence.
Written symbolically :
(1)
wher e (J' is the specular angle.
S p ecular glossm eter. An instrument for measur-
ing specular gloss at a fixed specular angle.
Geom etric conditions of i llumination and view.
Include the specular angle and the apertures of
the source and r eceiver .
S ource . A luminous body or area by which the
specimen is illuminated.
R eceiver. A photocell or an enclosure equipped
with phot ocells r esponding t o luminous flux
from the specimen. The sensitive surface of the
photocell or the opening to the enclosure is
called the receiver entrance window.
Plane of measurement . The plane passing thr'ough
t he axes of incidence and view.
A perture (of a source or receiver) . The plane angle
subtended b y the source or r eceiver en trance
window at the respective collimating lenses.
With rectangular apertures th e angles, ex and (3,
may be used t o describe the angular size of the
aperture in the plane and perp endicular to the
plane of measurement, r espectively . In com-
puting an aperture it is convenient to find the
tangent of tIle half angle of th e aper ture. For
example: the source 8-perture, lX , in the plan e of
measuremen t is equal to 2 tan- J(y/f), where y
is the half dimension of the source in the plane
of measurem ent, and f is th e focal length of the
collimating lens.
Fresnel reflector (nonmetallic). Specularly reflects
a fraction, Gs , of luminous flux computed from
the Fresnel equation [2],
G _ !!;. [sin 2 (i- r ) tan2 (i- r)]
.- 2 sin2 (i+ r ) +tan2 (i+ r) ,
(2)
587
 where E is the illuminance, i is the angle of PERFECT PERP ENDICULAR PERFECT
SOURCE RECEIVER
incidence, and r is the angle of refraction. COLLIMATOR COLLIMATOR

A more convenient form of this equation is ob- POINT

SOUR CE
POINT
RECEIVER
tained by expressing the angle of refraction in
terms of the angle of incidence and the index
of refraction, n, of the material. This is:

OPTI CALLY FLAT SURFACE

FIGURE 2. Theoretical glossmeter, employing point source

(3) and point receiver.

2. Practical Glossmeters
Perfect d~fJnsor. Reflects light according to
Lambert's Cosine Law, which states that the Most glossmeters in commercial use today em-
solid angular density of the luminous flux re- ploy an incandescent-filament lamp as a ligh t
flected from the surface in any direction varies source, a source lens to collimate the inciden t beam,
as th e cosine of the angle between that direc- a receiver lens to obtain at the receiver entrance
tion and the perpendicular to the surface. window an image of the source reflected from a
Glossiness. The lustrous appearance of a surface Fresnel reflector, and some type of photoelectric
ascribable to its gloss. receiver [4], (see fig. 3). The source aperture I S
Lightness. That appearance attribute of a speci-
PRACTICAL P E RPENOICULAR
men which correlates with its luminous diffuse SOURCE
PR ACTICA L
REC E IVER
CO LLIMATOR
reflectance. COLLIMATOR
,/
PRACTICAL
III. Glossmeters SO UR CE

Instruments for the measurement of specular

gloss have been used for more than 30 years. In
1937, however, Hunter [3] described a photoelec- OPTICALLY FLAT S URFACE

tric modification of the McNicholas goniophotom-
eter, which is the earliest known photoelectric
specular glossmeter. Since then glossmeters of
this type and similar photoelectric types have been
manufactured commercially.
1. Theoretical Glossmeter
The theoretical glossm eter consists of a so urcc,
:source and receiver collimators, and a receiver
(see fig. 2). If we had to deal only with an in-
finitesimal, un polarized light so urce of high inten-
'sity, an infinitesimal receiver of high response, and
perfect collimators, the problem of geometric tol-
erance specification would be eliminated. Such
hypothe tical glossmeter elements would have min-
imum aperatures and no lens aberrations and
,could be duplicated without causing error in the
gloss readings. Measurement of specular gloss
would then be completely reproducible from in-
:strument Lo instrument if the specular angles
were accurately determined.
FIGURE 3. P rac tical glossmeter, em ploying practical source
and practical receiver.
determined by the dimensions of the source and
the focal length of th e collimating lens. The re-
ceiver aperture is determined in a similar manner,
using the dimensions of the receiver entrance
window. Standard apertures may be specified,
but errors of duplication cannot be avoided.
The collimating lenses used in commercial
glossmeters are usually simple, uncorrected lenses,
subject to such lens errors as astigmatism, spheri-
cal and chromatic aberration, and coma. Cor-
rected lenses could be used, but the accuracy re-
q uired in mos t gloss measurements can be obtained
without this added expense. Because most com-
mercial glossmeters are compactly made, the dis-
tances between elements and the apertures are
physically small. This leads to errors in the set-
ting of the specular angle, measurement of the
source and receiver apertures, collimation of the
incident beam, and positioning of the focussed im-

.588 Journal of Research

______________________________________________ ~ ____ ~~~e~~ ___ ~.~~~~--.-- __ ~ _____

age of the source in the center of the r eceiver
entrance window when reflected from a flat
mirror surface.
3 . The Versatile Glossmeter
In order to investigate the prccision to which
the many geometric variablrs of a glossmeter
must be sp ccificd , an existing goniophotomcter
was modified to make a versatile glossmeter (see
fig . 4). This glossmeter has provision for ac-
curately controlling (1) the source and r eceiver
apertures in bo th the a and (3 planes; (2) the angles
of illumination and view; (3) the collimation; and
(4) the position of the source image relative to the
plane of the receiver aperture. Each of these
variables can be changed individually to any
desired degr ee of departure from the ASTM
specifications.
Control of th e source ap erture is accomplish cd
by employing a pair of cond ensing lenses to foc~s
an unage of the source on an adj ustable metallIc
aperture. This aperture, servin g as the n ew
source of controllable si - e, is located in the fo cal
plane of an achromatic collimating l ens, the focal
length of which is accurately known.
The specular angle is easily controlled since the
so urce and receiver are located on the hori, ontal
arms of a goniophotometer of which the angular
scale may be set to within ± 0.1 °, and the speci-
m en of which the gloss is to b e m easured is placed
vertically on the turntable of the goniophotometer.
The r eceiver collimating lens also is an achro-
matic lens for which the focal length is accurately
known. It is of such aperture and so positioned
that there is no vignetting of rays that should
reach the r eceiver. The receiver aperture is
precisely lo cated so that the image of the so urce
is brought to focus at the plane of this aperture.
Since there is no need to conserve space, this
glossmeter was made large enough to allow for
high accuracy of measurement of the apertures
and the distances between elements.
IV. Glossmeter Variables
A surface lacking the smoothness of polish ed
black glass will r eflect som e of the incident light
in directions adjacent to that of specular reflec-
tion; consequently, some of the reflected ligh t
may fail to reach the instrument receiver, and a
lower gloss reading mas: b e obtained. It should
Measurement ·of 60 ° Specular Gloss
883 124- 50- -3
INTEGRATING _ _'\\\
PERP ENOIC UlA R ENCLOSURE
LIGHT
SOURCE
VAR IABLE RECEIVER
APERTUR E
--.<:~*~ ~-- VARIABLE ANG LES
BLACK GL ASS STANDARD
FIG U RE 4. NBS versatile glossmetel', emplo ying variable
sou rce and receiver apertw'es, achl'omatic colli mators, and
variable angles of i llu mination and view.
bc pointed out that the specular r efl ectance, and
consequently the gloss, of a nonmetall ic specimen
is a function of the index of refraction of the
material of the specimen (see eq 3) as well as
of the smoothness of the surface. Polished black
glass of known index of refraction h as been used
to calibrate the upper end of the gloss scale.
For low-gloss finish es, such as are used for cam-
ouflage by the Armed Services, the plane surface
of a block of magnesium carbonate, which ap-
proxlinates the perfect diffuser, has b een used as a
standard at the lower end of the scale. This
leaves the question of what material to use for
interm ediate standards. Horning and Morse [5J
implied that in the medium- to low-gloss range
average paint specimens have reflected light-
flux distributions more like glazed ceramic tiles
than like depolished black glass. From t his
standpoint tile should be preferable for use as
gloss standards in the paint industry.
R ecent cooperative work with the Sherwin-
Williams Co. revealed, however, that better agree-
ment in the measurement of high-gloss paints h ad
been obtained on different glossmeters with the
use of depolished black glass standards than with
ceramic tile standards. In fact ceramic tile
standards produced much larger instrument dis-
crepancies for the same specimens. In order to
analyze this situation, goniophotometric curves of
fractional reflectance in the region of specular
reflection were obtained for the pain t specimen , a
tile standard, and a sand-blasted depolished black
glass standard having about the sam e 60° specular
gloss (80 on the ASTM scale). These curves are
shown in figure 5. For these m easurements the
goniophotomeLer SOUl'ee aperture was made 0.3°
circular, and the reeeiver aperture was mad e 0.6°
589
circular. A polished black glass standard was 100 o.

used , and the results are plo tted on a fractional

r eflectance scale times 1,000, thus giving a value
of 1,000 to the perfectly reflecting plane surface.
From these curves it is seen that the ligh t-flux 100

distribution of the sand-blasted black glass stand- o

o
Q
ard is very similar to that of this particular high- ~ -
gloss paint specimen, whereas the distribution of UJ
<.) I o.
Z
the tile standard is quite different. If a glossmeter <t
I-
<.)
having accurate ASTM apertures is used to deter- UJ
-'
IL / \
min e the gloss of this paint specimen , the reading UJ
'\
will be the same no matter which standard is used.
a:
-'
<t
I.
TILE STA N DARD
fl .....--GLOSSY BLAC K PAINT

Z
On the other hand, if the receiver aperture were o
>=
too large, very nearly the true gloss will be obtained Cl
<t
a:
with the glass standard, but a much lower reading h !jj ~
IL SANDS LASTED BLA CK
GLASS
I
will be obtained for this specimen if the tile
standard is used for calibrating the instrument. /I .~

The conclusion might be drawn at this point that

instrumental aperture variation might be mini- .0 I
./
~
'" 0- "'-
~
mized by using the glass standard, and this would --------
to
be true if all high -gloss specimens had flux distri- -12 B 6 -2 0
DEPARTURES FROM SPECULAR ANGLE, DEGREES
2 4 6 8 10 12

butions like th e one illustrated here. However ,

as implied by Horning and Morse, many specimens
are encountered with reflected light-flux distri-
butions more like that of the tile standard, in
which case nearly correct gloss readings would be
obtained with this standard even though the
receiver aperture were too large, whereas calibra-
tion with the glass standard would cause the
instrument to give a reading higher than the
correct value. Thus it is seen that if the refl ected
light-flux distributions of specimen and standard
are nearly identical, the size of the apertures is
not critical.
However, it is no t practical to determine the
distribution of each type of specimen , and it is
even less practical to attempt to duplicate these
distributions even approximately with durable
materials that migh t be used as standards. The
more practical approach to the problem is to make
the glossmeter apertures as accurate as possible so
that the instrument will give correct gloss r eadings
no matter what the reflected light distribution of
the specimen. Although glossmeter manufactur-
ers may claim that the apertures ar e exactly cor-
rect, it would be desirable for the instrument oper-
ator to have a simple method to check their ac-
curacy. This may be done by taking advantage
of the facts pointed out above, namely that a
glossmeter will r ead correctly two standards hav-
FIGURE 5. Luminous fractional reflectance curves of high-
gloss materials.
Angle of illumination and view, 60°, source aperture, 0.3 0 , Circular, re-
ceiver aperture O.Go.
ing radically differ ent distribu tions of reflected
light-flux only when apertures are correct.
In the usual method of glossmeter operation, a
polished glass surface is employed to make the
initial calibration. The reading subsequ ently ob-
tained for a tile standard should be the true value
if the apertures are correct, too large if the receiver
aperture is too large or too small if the aperture is
too small, other adjustments being correct. It
may be noted that the r eproducible plane surfaces
of different r efractive index employed by Moore
and Hunter [6] to avoid inaccuracies resulting fl'om
abrasions of a black glass standard do not serve
to disclose errors arising from nonstandard aper-
tures, In contrast to reflectance measurements
where the apertures of the illuminating and view-
ing beams are relatively unimportant, the aper-
tures employed in gloss measurement must be
carefully specified and held with close tolerance if
accurate glossimetry is to be accomplished. It
should be stated here that in con trast to previous
practice, it has been found no t sufficient to specify
the combined spreads of source and receiver aper-
ture but each must be specified separately because

590 Journal of Research

 ------ ----------------------------------------

the ratio of source aperture to r eceiver aper ture the receiver apertme. On the other hand, th e
affects significantly the measured gloss value. measmed values for th e depolished black glass
standards remain r easonably constant over the
1. Receiver Aperture wide range of ap ertmes employed. This is ascrib-
To illustrate th e magni.tude of the effect of able to the fact that high-gloss tile produces
variation of th e r ec tangular r eceiver aperture in narrow-angle scattering n ear the specular angle,
the plane in which th e sp ecular angle is measured , whereas depolished blac],:: glass produces 'wide-
designa ted ex, and in the plane in th e polar direc- angle scattering.
tion perp endicular to ex, designated {3, two series
100.----.-----------,----------,-----------,
of measnrement were made on plaques of different
gloss. This was done for both depolish ed (sand-
o------<>----~---..()...---<>-----<> G 90
blasted) black glass, and for glazed ceramic tiles.
The results of th ese measm em en ts are illustrated
~ ___ ~ G 80
graphically in figm es 6 and 7. These figures show 80 0-
=i= ---r -g--
T 80
•
the effect of variation in r eceiver aper ture in the
ex and {3 planes, respectively, on the m easm ed ~ T 70
gloss values of tile and depolish ed black glass ij::.: ~
g ~ ~ G 70
(j)

standard . It will be no ted that the m easured L?

Z

- __-------<0------- ~ ~g
values for th e tiles arc a pronounced function of ~ 60 0- :;---- -<;----T--=i=--- -g
w
cr
100.----.-----------.-----------.----------,

G 90
T 80
cr
w
~
w
::;:
-
o - ___ -o- ___ -o- ___ -CJ- _ _ -O- _ _ _ _<>
T 50
G 50

(j) 40
T 70 (j)
o
~
80 L?

(j)
L?
Z 20
o
<l
W 60
cr
cr
w
~
w
::;:
0L---74~.4~X~8::~::::4;.4;:X;10;=~==::4~.4~X~,2~~~--~
(j)
T30 RECEIVER APERTURE , DEGREES
(j) 40
o
~
L?
F IG U H E 7. Glossm eter Te adings obtained by varying the
di mension of the receiveT a pertw 'e in the {3 plane.
T 20 T he depolished black glass specimens are designa ted (I , fo llowed by t he
nominal gloss value; the glazed ceramic t ile specimens are designated T, followecl
by tbe nominal gloss valne.
20
2. Source Aperture

To determine the effect of somce aperture

variation on glossmeter r eadings and thu s be able
O~--~----------~--------~--------~
to determine permissible tolerances, data were
2.11 .7 4.11.7 6xll .7 obtained on the variation of m easm ed gloss with
RECEIVER APERTURE, DEGREE S
source apertme with th e r eceiver apertme held
FIG U H E 6. GlossmeteT readin gs obtained by varying the constan t a t th e value prescribed in the ASTM
dim en sion of the receiver ap erl1tre in the c< plane. m ethod mentioned earlier. It was found that the
T he depolish ed black speci mens are designated G, followed b y the nominal variation of somce apertme in th e ex plane affected
gloss value; the glazed ceramic tile specimens are designated T , fo llowed by
tbe nomin al gloss value. the measllred gloss of the plaques as shown in

Measurement of 60 ° Specular Gloss 591

figure 8; variation in this plane did not produce
as pronounced measured gloss differences as for
the receiver aperture variation. Variation in the
(3 plane had a negligible effect on th e measured
gloss values, so the data are no t shown here.
1 00 r--------r------~------r_----_,------_r~
<>---<>---- -o-- - - -~--- ~
a sharp focus, the image will be surrounded by a
red haze. The effect of position of the source
aperture with respect to th e receiver entrance
window was explored on th e exp erimen tal gloss-
meter by placing spectacle-type lenses of low power
between the specimen and receiver collimating
len s. This procedure was used because th e colli-
mating lenses are not easily moved from their
G 90
positions of optimum collima tion. By using posi-
t.ive or n egative lenses the image of the source was
obtained before or behind th e plane of the r eceiver

~ :::
80
entrance window. Figure 9 illustrates the effect
produced on the measured gloss values. It shows
that th e greatest effect is produced on th e glass
If)
<!l
values of the tile plaques, and that the effect on
~ 60 measured gloss values is no t negligible for tile but
a -.() G 60
"'o--=~'i"~.::-:.:-:..:-<>-~.;:.:==::-<>-~::_
«
w -'T 60 is practically so for depolished glass. The data
0: also show that there is no preferred direction of
••--~.~--~~------.----~~ T5 0
0:
W o---~----<>------~---_o G 50
f-
W 100 r----.-----r----.-----~--~----~----~
::;: 40
If)
• • T 40
If)
o
-' 0------0--------<> G 90
<!l
• • T 30

80 e- ~
-<> G 80
.. T 80
20 • • T 20

• T 10
(j)
<!l
Z
--
0- - ~~ ~ G 70
T 70

o
• • TI <l
w
60
O ~------~----~----~~----~----~~
I x 1.3
SOURCE
2xl.3 3 x l.3
APERTUR E , DEGREE S
4 xl.3
cr:
cr:
W
f-
- ..... T 50

W
0------<>---------0 G 50
FIGURE-:'" 8. Glossmetel' readings obtained by varying the ::i'
dimension of the source aperture in the a plane. ~ 40 - T 40
..-----------------....
The depoJisbed black glass specimens are designated 0, followed by the o
-'
nominal gloss value; t he glazed ceramic tile specimen s are designated T, <!l
followed by the nominal gloss value. • ---------------~
... ..... T 30

3 . Position of Source Image

20 ••- - - - - - - - - - - - - - - - . T 20
Another question of practical importance is the
effect of the position of the image of the source. A
....- - - - - - - - - - - - - - - - . TI O
properly adj ust.ed glossmeter will form an image
of the source centered in the receiver en tran ce ••------••--------_.. TI
O~--~----~----L---~-- __~____L__ _~
window when a fiat polished s urface is placed in 20 10 0 10 20 30
the specimen position. Commercial glossmeters BEFORE BEHIN D
IMAGE DISTANCE FROM RECEIVER APERTURE PLANE, em
using simple lenses are difficult to adjust because
chromatic abberration of the lenses causes light of FIGU RE 9. Glossmetel' readings obtained by va1'ying the
different wavelengths to focus at different posi- position of the focus ed i mage of the source relative to plane
tions. If the red rays are brought to a sharp of the receiver aperture.
focus, the source image is s urrounded by a blue The depoJished black glass specimen s are designated G, followed by th e
nominal gloss value; the glazed ceramic tile specimens are designat ed T,
haze. Conversely, if the blue rays arc brought to followed by the nominal gloss value. F ocal len gth of receiver lens is 54.6 em .

592 Journal of Research

----------.- ~--
departure from the plane of the receiver entrance whereas the tile plaques show s light decrease in
window. relative gloss with in creasing angle of incidence .
4 . Specular Angle
1oor-,-----,----.-----.----,-----,----.--.
Glossmeters are usually adjusted so that the
image of the source reflected from a plane reflector, 0------<>------,,-----0------0 G 90
such as that of polished opaque glass, falls on the
center of the receiver entran ce window. Suppose,
80
however, that in making this adjustment the angle "
of incidence is disturbed and that it. is found to be
If)
59 0 or 61 0 instead of 60 0 • What effect will this <!l ~.____~.~--~~==~-=~-~~G70
Z 0-- -0- :!: -- - • T 70
error in the specular angle have on the gloss o
<t
readings? It can be shown from the Fresnel equa - ~ 60 -~- --<> G 60
tion (see definition of Fresnel r efl ector) that a • • T 60
cr
departure of l a in specular angle at 60 0 will cause a LU ··----~·~---+·----~.~--_o. T50
f-
change of approximately 5 percent in the specu- LU -<>- ___ -0 G 50
::E 0 - - - - - --0--- ---0-----
larly reflected flux from a polished surface of re-
fractive index 1.53. In the usual method of oper-
If)

~ 40
.J
.~--_4
.~--~-- __ ~--__.
. T40
ation, the instrument is set to read correctly the <!l

value of the working standard , then the gloss of

the test specimen is read relative to it. If the
•.-----o.~--~----~---. T 30

standard and specimen have nearly the same sur-

face characteristics, no appreciable error win r esult.
Figure 10 shows the effect on the measured gloss
of both tile and depolished glass standards with
departure from the 60 0 specular angle. The
illuminance was adjusted for each angle so that the
instrument indicated the 60 0 specular gloss of the
polished glass reference standard . It will be noted
from the figure that for the range of angles in-
vestigated the measured gloss of the depolished
glass plaques increases with increasing specular
angle, bu t the measured glos of the tile plaques
decreases slightly. In making th e adj ustment for
the standard, the illuminance for 59 0 , as an ex-
ample, had to be increased. Then two factors are
operating, (1) the increased illuminance and (2)
the deCI'eased directionalrefiectance du e to Fresnel
reflection. The effect on the depolished glass
plaques is explainable by the assumption that the
second factor acts more importantly than the fJrst,
since the surface is composed of many facets in the
gross plane of the plaque. The eff ec~ on the tile
plaques is explainable by the assumption that the
second factor is operating less importantly than
the first, since the higher index of refraction of this
material causes a smaller r elative increase in r e-
flectance with increasing specular angle . Since
the instrument is caused to give the same reading
for the reference standard the depolished black
glass plaques show slight in crease in relative gloss,
20 ••- - - -.....---- -- . - - - -.....- ---.. T 20
••----~----...- - - -.......--- -- . T 10
----~-------___- - _ . T I
O ~~----~--~----~----~----~--~--.J
57 58 59 60 61 62
SPECULAR ANG LE , D E GREES
FIG U RE 10. Gloss m eter readings obtained by varying the
specular angle.
The depoJished black glass specimens a re designated 0, followed by th e
nominal gloss value; the glazed cCt'amic tile specimens a l'O des iglUlted T,
followed by the nominal gloss value.
V. Diffuse Correction
1. Simple Theory
We come now to the problem of diffuse correc-
tion. Figure 1 shows that the reflected light re-
ceived by the glossmeter may be considered to be
composed of two parts, the specular reflection or
true gloss contributed by the surface of the speci-
men and the diffuse reflection contributed by the
body of the specimen . Thus we see that the light-
scattering property of the body of the specimen ,
or in a sense its body color, affects the measured
gloss. In the case of high-gloss specimens, the
diffu se component is a small fraction of the instru-
ment reading, and many laboratories working with
high-gloss materials do not correct or allow for it.

Measurement of 60 0 Specular Gloss 593

In the case of low-gloss specimens, however, this The solid angle subtended by this receiver is
correction should be made especially when com-
paring specimens of different lightness.
(a) Hemisphere Method

Picture a perfectly reflecting, perfectly diffusing
specimen which reflects no light specularly but
which causes a "gloss" reading on the instrument
because some of the diffusely reflected light reaches
the receiver. What should be the magnitude of
the correction to be substracted from the instru-
ment reading to obtain the true gloss of the speci-
men? Lambel't [7] showed that a perfectly diffus-
ing specimen reflects light in such a manner that
the luminous intensity in any direction is propor-
tional to the cosine of the angle, ..p, between the
perpendicular to the surface and the direction of
reflection. He also showed that the perfect diffuser
would appear equally bright in all directions, that
is, that the luminance is constant.
Let us assume that the luminous flux reemitted
by a perfect diffusor having luminance, B, and
area, dA falls normally on the surface of a hemi-
sphere of radius, R (see fig. 11). First we wish to
compute the total flux, F, falling upon the entire
hemisphere. The flux, dF, in a direction making
an angle ..p with the perpendicular to the diffusor
is evidently proportional to the projection of dA
on a plane perpendicular to the direction ..p, or,
dA cos ..p, and to the solid angle, dQ, sub tended at
the diffusor. The proportionality constant is the
luminance, B. Thus
where ex is the aperture angle in the plane of
measurement (the plane in which the specular
angle (J' is measured), and {3 is the aperture angle
in the plane perpendicular to the plane of measure-
ment. The angles ex and {3 are small and are
measured in radians. Then
f = Ba{3 cos (J'dA. (6)
Thus, of the reemitted luminous flux, the fraction,
F r , entering the receiver aperture is
f a{3 cos (J'
Fr F 7r
(7)
For the ASTM receiver,
Thus Fr is computed to be 0.0025. This means
that the perfectly reflecting, perfectly diffusing
specimen although in reality having no specular
gloss would cause the instrument to read 2.5 if a
perfect plane reflector were to read 1,000.

dF= B cos ..pdAdQ. (4)

The solid angle,

where dQ is now the hollow cone sub tended at the

specimen by the zone of width Rd..p, whose semi-
SPECIMEN
angle is..p. Therefore,
FIGURE 11. Diffus e correction for the peljectly diffusing
{7r/2 surface treated by the hemisphere m ethod on the basis of
F = 7rBdA) o 2 cos ..p sin ..pdl/;= 7rBdA. (5) simple theory.

(b) Tangent Sphere Method

N ow consider the flux,!, which falls on the receiver
at an angle (J'. The magnitude of the diffuse correction may be
From eq 4 above obtained very simply by considering the illumi-
nance E at any point on a sphere of diameter D
df= B cos (J'dAdw. tangent to the diffusor (see fig. 12). Let the

594 Journal of Research

 This is the same as eq 7 obtained by integration of
luminous flux over the hemisphere surface. It
should be noted that the diffuse correction is the
ratio of the area of the tangent sph ere intercepted
by a and .6 to the area of the whole sph er e. It
may be said, therefore, that the diffuse correction
is numerically equal to the fraction of the tangent
sphere intercepted by the solid angle form ed by
the instrument aperture.
2 . Expansion of The ory

Practical surfaces, neither perfectly reflecting

nor perfectly diffusing, will generally have smaller
corrections than 2.5, but, except for black sur-
faces, the correction is appreciable and must be
made in the low-gloss range in order to grade cor-
rectly low-gloss surfaces of different luminou s re-
flectance. To determine the diffuse corrcction one
must take into account the following: (a) rcflec-
SPECI M EN
tance of the specimen, (b) departure of practical
FIGURE 12. Diffuse correction for the pelJectly diffusing surfaces from perfect. diffusors, and (c) polarization
surface treated by the tangent sphere m ethod on the basis of the source.
of simple theory.
(a) Specimen Reflectance
luminous intensity in any direction perpendicular
to the pcrfect diffusor of area dA be represented by With a goniophotometer or other glossmeter in
IodA and th e intensity in other directions by which the illuminator can be rotated about the
I",dA. Then , specimen in the plane of m eas uremcnt, this cor-
rection is most easily evaluated by illuminating
the specimen perpendicularly and viewing at an
angle of 60°. When this is clone the area illumi-
The illuminance, E"" at any point on the spher e is nated is one-half of that previously illuminated,
the intensity in that direction multiplied by the but the flux den ity is twice as great. If t he sur-
cosine of the angle of incidence and divided by the face is uniform, the diffuse correction is obtained
square of the distance, D 2cos 2if;, from the perfect directly exccpt that a very small error is in volved
diffusor to that point on the sphere. Therefore, because the surface reflection is lcss for perpendic-
ular than for 60° illumination. For a polished
I", cos if;dA I odA specimen having a refractive index of 1.53, the
D 2 cos 2 -if; D 2· (8) surface reflectance at 60° illumination is about
9 percent and at 0° is about 4 percent. This means
Thus we see that the illuminance on the sphere is that an additional 5 percent is added to the body
independent of the angle of view. Therefore the illuminance, but since this flux is di stributed over
ratio, F of the flux on a receiver to the total flux
T, a hemisphere the increase in light enLering the
is simply the ratio of the area, a, of the receiver receiver is small. The fraction of the total light
projected on the sphere to the area, A', of the entering the receiver was previously computed to
whole sphere. But a, located at an angle 8' from b e 0.0025. Since the surface-reflected light is
the perpendicular to the diffusor, is a6D2cos8' now decreased by 5 percent, the body-reflected
and A' is 7r D 2. Therefore , light is increased by the same fraction, so that the
diffuse correction for 60°, - 60 ° condition as m eas-
a a{3D2 cos 0' a{3 cos 8' ured under 0°, - 60° conditions is high by 5 p ercent
F T = A' 7rD 2 7r (9) of the magnitude. This means that for a speci-

Measurement of 60° Specular Gloss 595

men having a measured diffuse correction of 0.0021 It will be recalled that the reflectance of mag-
the true diffuse cOTI'ection would be 0.0020 or, on nesium oxide prepared in a prescribed mann er [10]
the usual specular gloss scale, the discrepancy is assigned a value of 1.00 when illuminated at
would be 0.1 gloss unit. This discrepancy in the 45° and viewed perpendicularly. For 0°, 60° con-
diffuse correction is proportional to the gloss, and ditions the directional reflectance of these oxide
is zero for a perfectly diffusing surface. In prac- specimens was found to be only 92.5 percent of
tice this discrepancy is so small that correction for the value assigned for 0°, 45° illuminating and
It IS unnecessary. If, on the other hand, 60° illu- viewing conditions. This accounts for nearly half
mination and perpendicular viewing are used, the the discrepancy found between the computed and
instrument reading will have to be halved to obtain the measured values for the diffuse correction.
the diffuse correction, because the proj ected area
(c) Source Polarization
for perpendicular view is twice that for 60 ° view.
(b) Pradica l Diffusors It will be recalled that the Fresnel r eflectance
of a surface is a function of the degree of polariza-
On the master instrument, figure 4, at this tion of the incident light as well as the angle of
Bureau it was found that magnesium oxide, which incidence and the index of refraction of the sur-
is often thought of as the practical approach to face. The degree of polarization of the source
the perfectly r eflecting, if not perfectly diffusing was measured by rotating a polaroid filter in the
surface, has a diffuse correction of 2.1. This emitted beam and r ecording the maximum and
value was obtained numerous times and though minimum transmittances. The difference of the
not vastly differen t from 2.5 certainly gives a larger readings was divided by the sum to obtain polari-
discrepancy than could be accounted for by ex- zation. The data obtained show that this source
perimental error. Previous investigation of the (a C- 6 filam ent) is polarized 5 percent along the
diffusing properties of the surfaces of magnesium axis of the coil, or perpendicular to the plane of
oxide by McNicholas [8] and Preston [9] revealed measurement. Filament polarizations of 15 to 20
that it is not a perfect diffusor. This was con- percent for straight filaments were reported by
firmed by us for magnesium oxide; and similar Worthing [11]. Because the 60° angle is almost
data were obtained for two other materials. the polarizing angle for glass, approximately 5
Figure 13 illustrates graphicall y the results of this percent less incident flux is required to obtain the
investigation. same reflected flux from a polished glass surface
120r---.---- -- .- -- - - - . - - -__- .____- , than if the source were unpolarized. Although
Prac ti cal Dif f us i ng
the polarization is d estroyed by a low-gloss sur-
Surfoces face, partial polarization of the incident light
f-
Z
causes a greater fraction to be reflected from the
w
u 100 polished surface of the high-gloss standard. Con-
'"
w
n. sequently, the reduced incident light on the speci-
W men reduces the computed value of the diffuse
u
z
..: component by 5 percent of its magnitude.
f-
U
w
80 These three sources of error, (1) surface reflec-
oJ
"-
W
tance, (2) departure from perfect diffusor, and (3)
'" polarization of the source, taken together account
oJ
..:
z for the discrepancy between the computed and
0
~ 60 the measured values of the diffuse correction.
u
w
'" ° MgD
0
() Vi tral i te
VI. Gloss Scale and Standards
• Por ce la i n Enam e l Any assembly of apparatus such as a glossmeter
40 requires calibrating and the adoption of a scale on
30 45 60 75 90
ANGLE OF VIEW, DEGREES which measurements will be based. The theo-
FIGURE 13. D eparture of practical sU1jaces from the con- retical specular gloss standard is the perfect mirror
cept of the perfect dijJuSOI·. that is assigned a value of 1,000. The practical

596 Journal of Research

primary standard is a piece of polish ed black glass
whose specular r eflectan ce is computed from
Fresnel's equation when the index of refraction of
the glass i known. Black glass is used for this
purpose, because all of the light flux not specularly
reflected is absorbed by the body of the material
and the problems associated with diffuse correction
are eliminated.
In order to providc means for ch ecking the ad-
justment of glossmeters and to assist in more
accurate m easurement of 60° specular gloss, this
Bureau has investigated the suitability of various
materials for use as standards, and has r ecently
assembled sets of standards for issue. Each set
consists of two white Vitrolite gloss plaqu es havin g
nominal gloss values 1 and 90, and eigh t glazed
ceramic plaques having nominal gloss values 10, 20,
30,40, 50, 60, 70, and 80. The plaques ar e approxi-
mately 4}~ in. square. Figure 14 is a photograph
of the plaques together with the box in which the
set of ten standards is furnished. Although the
plaques are not intended to be used as standards
of distinctness-of-image gloss, the photograph in
figure 15 h as been taken to indicate approximately
the gloss steps involved, since for the e plaques 60 ° FIG URE 14. Set of NBS gloss standards, with box in which
specular gloss steps agree in order with distinct- they a1'e issued.
ness-of-image gloss steps estimates.
wher eas matte surfaces can best be differentiated
VII. Summary and Recommendations at large angles from the perpendicular.
Problems inherent in practical glos mcters have
Although photoelectric m easurem ents of spec- been empirically treated h ere. After permissible
ular gloss at a particular angle are convenient for discrepancies in gloss values have been decided
the evaluation of surface appearance, such meas- upon, the toleran ces for the geom etric variables for
urement do not completely evaluate the gloss 60° specular gloss can be obtained from th e curves
characteristics, much less the over-all appearance in figures 5 to 9. If a deviation of ± 1 gloss unit is
of th e surface. In general, the specular gloss of sp ecified, then the toleran ces can be readily ascer-
high-gloss surfaces can best be differentiated by tained. On th e basis of the data obtained by th e
observation at small angles from the perpendicular, authors on the versatile glossmeter and a pcr'mis-

FIGURE 15. Photographic representation of specular gloss standards.

The standards are arranged in order o[ decreasi.ng specular gloss [rom left to right, as shown by the more readily photographed gradation of the distinctness
o[ reflected images. For these standards distinctness-o['refiected-image gloss correlates with specular gloss.

Measurement of 60 ° Specular Gloss 597

883124-50--4
sible deviation of ± 1 gloss unit, the recommenda-
tions shown in table 1 are made.
T ABLE 1. Recommended tolerances for practical 60° gloss-
meters
Variable Recomm ended tolerance
Receiver aperture:
In plane or measurement ____ ___ ±O.l°.
P erpendicular to plane of meas· ±O.2°.
Ul·cment.
Source aperture in plane of meas· ±O.2°.
urement.
Position of source image , . ......... ±5% of the focal length of rc·
ceiver lens.
Specular augle . . .... . .... . . .... .. . . ±O.3° from the speci fi ed spec·
nlar angle.
, For uncorrected lenses, source im age should be focussed for the circle of
least confusion.
In view of the fact that these discrepancies are
cumulative, it is recommended that manufacturers
and users of practical glossmeters be careful that
the geometric variables be held well within the
recommended tolerances; for if all the variables
happen to contribute errors in the same direction
though they are just within the recommended tol-
erances, the r eading may be in error by as much as
5 gloss units.
The authors express their appreciation to D. B.
Judd for invaluable suggestions and advice during
preparation of this paper and also to E. M. Haga-
man for his ingenuity and assistance in the con-
struction of the versatile glossmeter.
VIII. References
[1] R. S. Hunter and D. B. Judd, ASTM Bulletin No. 97,
11 (March 1939) .
[2] A. Fresnel, Calcul des teintes que polarization devel-
oppe dans lames cristallisees, Ann. chim. phys.
17,194,312 (1821) .
[3] R. S. Hunter, Methods of determining gloss, J. Re-
search NBS 18, 19 (1937) .
[4] ASTM tentative method of test for 60° specular
gloss of paint fini shes, ASTM designation D523,
Book of ASTM Standards, Part II (1949) .
[5] S. C. Horning and M. P. Morse, The measurement of
the gloss of paint panels, Official Digest (Fed.
Paint and Varnish Products Clubs) 266,153 (1947) .
[6] D. G. Moore and R. S. Hunter, Use of liquid surfaces
as standards of specular gloss, J. Am. Ceramic
Soc. 24, 167 (1941).
[7] J. H. Lambert, Photometria sive de mensura et
gradibus luminis, colorum et umbrae, p. 324,
Augsburg (1760) .
[8] H. J. McNicholas, Absolute methods in refl ectometry,
BS J. Research 1, 29 (1928) RP3.
[9] J. S. Preston, The reflection factor of magnesium
oxide, Trans. Opt. Soc. (London ) 31, 15 (1929) .
[10] NBS L etter Circular LC547, Preparation and color-
imetric properties of magnesium-oxide reflectance
standard (1939) .
[11] A. G. Worthing, Deviation from Lambert's law, and
polarization of light emitted by incandescent
tungsten, tantalum, and molybdenum , and changes
in the optical constants of tungsten with tempera-
t ure, J . Opt. Soc. Am. 13, 635 (1926) .
WASHINGTON, February 13, 1950.

598 Journal 01 Research